DEV Community

venkat sathu
venkat sathu

Posted on

How we built ThreatLedger — AI-powered NDR on AWS Aurora + Vercel in 72 hours

This post was created for the purposes of entering the H0: Hack the Zero Stack Hackathon. #H0Hackathon

The problem

Small businesses face the same network threats as enterprises but can't afford $100K/year NDR tools. We built ThreatLedger to change that.

What we built

ThreatLedger is a cloud-native Network Detection and Response dashboard that turns raw security logs into actionable threat intelligence using AI.

Live demo: https://threatledger-app.vercel.app

The stack

  • Frontend: Next.js 16 on Vercel (scaffolded with v0)
  • Database: Amazon Aurora PostgreSQL (AWS us-east-1) with pgvector
  • ORM: Prisma
  • Auth: Clerk
  • AI: ChatGPT API

How it works

  1. Upload Suricata, Zeek, Firewall CSV, or AWS VPC Flow logs
  2. Custom parsers extract and normalize alerts
  3. Correlation engine groups alerts into attack campaigns with composite risk scores
  4. Kill chain mapping shows attack progression
  5. Claude API generates plain-English threat summaries

Why Aurora PostgreSQL

Aurora gave us a production-ready database from day one. We enabled pgvector for future semantic search across 21,742 IP reputation records. The Prisma + Aurora combination gave us type-safe queries with zero configuration overhead.

Biggest challenge

Getting Prisma 7 working with Aurora on Vercel's serverless environment required configuring PrismaPg adapter with connection pooling and handling SSL correctly.

What's next

Semantic similarity search using pgvector, MITRE ATT&CK mapping, and real-time log streaming.


Built for the H0: Hack the Zero Stack Hackathon #H0Hackathon

Top comments (0)